Effect of changing the rare earth cation type on the structure
and crystallization behavior of an aluminoborosilicate glass
A. Quintas, D. Caurant, O. Majérus
Laboratoire de Chimie de la Matière Condensée de Paris, UMR 7574, ENSCP, Paris, France
A. Quintas, J-L. Dussossoy
Commissariat à l’Energie Atomique, Centre d’étude de la Vallée du Rhône,
DEN/DTCD/SCDV/LEBV, Bagnols-sur-Cèze, France
T. Charpentier
CEA Saclay, Laboratoire de Structure et Dynamique par Résonance Magnétique,
DSM/DRECAM/SCM – CEA CNRS URA 331, Gif-sur-Yvette, France
An aluminoborosilicate glass, containing high amount of rare earth (RE) accordingly to the following
composition 50.68 SiO2 – 4.25 Al2O3 - 8.50 B2O3 – 12.19 Na2O – 4.84 CaO – 3.19 ZrO2 – 16.35 RE2O3
(wt.%), is currently under study for the immobilization of nuclear waste solutions. In this work, we wanted
to investigate the effect of changing the RE cation type on the glass structure and on its crystallization
behavior. For this purpose, a glass series was elaborated in which the nature of the RE is varying from
lanthanum to lutetium. In this glass series, only little effect was observed on the glass structure. On the
contrary, a strong impact was put in evidence on the crystallization behavior through different heat
treatments. A slow cooling of the melt at 1°C/min, revealed significant crystallization of apatite
Ca2RE8(SiO4)6O2 in sample containing rare earths with ionic radii close to that of calcium. Another heat
treatment consisting of successive nucleation and growth stages, performed to force the crystallization in
the bulk and reduce any surface crystallization effect, put in evidence the existence of a strongly
heterogeneous second rare earth rich silicate phase for samples containing RE with low ionic radius (from
Y to Lu).
Introduction
In order to immobilize radioactive wastes, stemming from the reprocessing of high
discharge burn up nuclear fuel, a new confinement glass matrix has been recently envisaged.
A rare-earth (RE) rich glass has been selected for its good performances (melting at 1300°C,
good chemical durability and waste capacity, low crystallization tendency) [1,2]. A simplified
seven oxides glass model composition (referred to as glass A): 61.81 SiO2 - 3.05 Al2O3 - 8.94
B2O3 - 14.41 Na2O - 6.33 CaO - 1.90 ZrO2 - 3.56 RE2O3 (mol.%) was adopted to facilitate
crystallization studies and structural investigations. In previous studies [3-5] of glass A, only
Nd or La were used as rare earth in the glass composition. Heat treatments performed on
both rare earth glasses revealed significant differences between Nd and La rich glasses in the
crystallization of an apatite phase (a much lower crystallization rate was obtained with the La
bearing glasses) which was the first demonstration of the non-equivalent role of Nd and La
towards the crystallization tendency.
In this work, we investigate how the nature of the rare earth may influence the glass
structure and its crystallization behavior. For this purpose, an extended series of ten glasses
has been prepared in which the nature of the trivalent RE is varied from lanthanum to
lutetium. The following lanthanides, classified in the order of decreasing ionic radius and
increasing field strength, were independently incorporated in the glass composition: La, Pr,
Nd, Sm, Eu, Gd, Er, Yb, Lu. In addition, a glass containing diamagnetic yttrium, which is a
trivalent rare earth that does not belong to the previous lanthanide series, was also prepared
(in terms of ionic radius, Y3+ inserts between Gd3+ and Er3+). All of these glasses (except the
Nd bearing glasses) contain 0.15 mol.% of Nd2O3 introduced as a local structural probe. The
influence of the RE cation type on the glass structure was studied by the mean of Raman
and MAS NMR.
Results and discussion
Both MAS NMR and Raman spectroscopies did not reveal significant effect of a
change of the rare earth cation type on the glass structure. Only a slight effect on the boron
speciation was noticed. According to 11B MAS NMR (figure 1) spectroscopy, the
[BO4]/[BO3] ratio decreases with increasing rare earth cationic field strength (the proportion
of fourfold boron species among total boron species varies from 41.6% for La to 32.9% for
Lu).
Figure 1 : 11B MAS NMR spectra recorded for the diamagnetic RE (La, Y and Lu) and
normalized to the same maximum BO4 intensity (B0=11.75T, νrot=12.5kHz).
In the meantime, crystallization studies were intended to follow the impact of the
nature of the RE on the crystallization behavior of the glass melt. Indeed, a high
crystallization tendency during the cooling of the melt is considered as a drawback for the
specific application envisaged here. The incorporation of α-radionuclides in crystals may
induce amorphization of the crystalline phases and the resulting swelling could lead to waste
form fracturation that can affect the long term durability. Two different heat treatments
were performed to study crystallization:
RG
100µm
Figure 2 : X-ray diffraction patterns of
the RE glass series restricted from La to
Gd after slow cooling of the melt at
1°C/min (λCoKα1=1.789Å).
Figure 3 : Back-scattered SEM image of
the neodymium bearing glass showing
silicate apatite crystals with different
microstructures formed during slow
cooling of the melt (1°C/min); RG:
residual glass.
- A slow cooling of the melt (1°C/min from 1350°C to room temperature) was done in
order to simulate the natural cooling in the bulk of nuclear waste glass containers after
casting and to determine the nature and the amount of the crystalline phases that may form
during slow cooling. It appears that crystallization is strongly affected by the nature of the
RE present in the glass composition. According to the diffractogramms presented in figure
2, glasses containing RE with smaller ionic radius than that of Eu do not lead to
crystallization of apatite Ca2RE8(SiO4)6O2 phase (figure 3). La-bearing glass also displays very
low crystallization tendency.
- Another heat treatment consisting of successive nucleation (2 hours at Tg+20°C) and
growth (30 hours at 934°C) stages was intended to force the crystallization in the bulk and to
reduce any surface crystallization effect. Compared to the slow cooling, this heat treatment
gives way to the crystallization of apatite crystals from La to Gd. Note that the progressive
shift of the diffraction peaks towards higher angle from La to Gd (i.e. increase of the cell
dimensions) is consistent with the decrease of the ionic radius of the rare earth. As for the
slow cooling heat treatment, a maximum crystallization rate is obtained for glasses
containing Pr or Nd. Moreover, another crystalline phase (not detected by XRD on massive
sample) can be observed from yttrium to lutetium and seems to grow at the expense of the
apatite phase. The nucleation process of this so-called P-phase is strongly heterogeneous as
it only appears on a very thin surface layer. The extent of crystallization of the P-phase was
enhanced by performing the nucleation and growth heat treatment in the same conditions
on powder glass samples (particle size: 80-125µm). The diffractogramms of the powder
samples exhibit peaks of the P-phase already observed in previous studies in glasses derived
from glass A and containing high amount of calcium substituting for sodium. This crystalline
phase was then identified as a calcium and neodymium silicate of composition Ca 13xNd2xSiO3 (where x≈0.18) determined by electron probe microanalysis.
In order to account for the maximum crystallization observed for Pr and Nd glass
samples, we suggest a thermodynamic approach based on the comparison of ionic radius
between RE3+ and Ca2+. In the apatite structure, RE3+ and Ca2+ may occupy the two 4f and
6h sites (respectively 9-fold and 7-fold coordinated). During the nucleation stage, the apatite
nuclei form at relatively low temperature (Tg+20°C). So the diffusion process is slow and as
the initial glass system is disordered, it is assumed that the nuclei formed during the first
stage of nucleation are disordered (random distribution of RE3+ and Ca2+ cations in both 4f
and 6h sites). In the slow cooling heat treatment, the nuclei form at quite high temperature.
In this case, we suppose that the first nuclei are also disordered in order to lower the entropy
difference between liquid and nucleus (-TΔS). The enthalpy of the first disordered nucleus is
lower in the case of similar ionic radii (fewer local lattice distortion) than in the case of
dissimilar ionic radii. Subsequently, at a given temperature, when the RE ionic radius is close
to that of calcium, the difference in free energy between the liquid and the crystal is greater
which lead to a greater crystallization driving force.
Conclusion
In conclusion, the glass network is only weakly affected by a change of the rare earth
cation type. On the contrary, the crystallization behavior dramatically depends on the nature
of the rare earth introduced in the glass composition. During slow cooling of the melt, the
formation of an apatite crystalline phase is strongly favored when the ionic radius of the
incorporated rare earth is close to that of Ca2+ (Pr, Nd or Sm) which can be accounted for by
thermodynamic arguments. An isotherm heat treatment enhances the global crystallization
tendency. With this last heat treatment, a new surface crystalline phase appears for glasses
with low apatite crystallization (glasses containing low RE ionic radius: from Y to Lu).
[1] I. Bardez, D. Caurant, P. Loiseau, J-L. Dussossoy, C. Gervais, F. Ribot, D. R. Neuville, N.
Baffier, Phys. Chem. Glasses 46, 3120-329 (2006).
[2] I. Bardez, D. Caurant, J-L. Dussossoy, P. Loiseau, C. Gervais, F. Ribot, D.R. Neuville, N.
Baffier, C. Fillet, Nucl. Sci. Eng. 153, 272-284 (2006).
[3] A. Quintas, O. Majérus, D. Caurant, J-L. Dussossoy and P. Vermaut, J. Am. Ceram. Soc. 90,
712-719 (2007).
[4] A. Quintas, O. Majérus, M. Lenoir, D. Caurant, K. Klementiev, A. Webb, J-L. Dussossoy, J.
Non-Cryst. Solids 354, 98-104 (2008).
[5] A. Quintas, T. Charpentier, O. Majérus, D. Caurant, J-L. Dussossoy Appl. Magn. Reson. 32,
613-634 (2007).